Webcasts
Abstract
The IBM Sterling Order Management team conducted a panel discussion in preparation for the 2021 Holiday Peak season on the IBM Order Management SaaS platform.
Thursday, September 16, 2021 @ 10:00AM | 90 Minutes | Eastern Daylight Time (New York, GMT-04:00)
Content
Missed it? No sweat! IBM has you covered. Below is all of the information. Please Share it with your colleagues.
As promised for those who were able to attend, the panel discussion replay, the slide deck, the Q&As, and URLs are posted below. We hope you enjoyed the technical discussion we had and were able to generate a full checklist to ensure you are prepared for the 2021 Holiday Season.
Speakers
With decades of experience and the willingness to ensure clients are successful this holiday season, some of IBMs best and brightest shared information for you to review. The IBM Order Management Support and Development team discussed some of the key configurations, features and best practices to help with peak holiday season traffic on the Order Management on Cloud platform. IBM discussed the key topics such as:
- Sourcing and scheduling setup and discuss performance considerations for the peak.
- Considerations for Web Store, and Peak Testing
- Self-Service Tool Demo wherein we highlight tools and techniques to understand key aspects of the service
- Meeting the SLAs for BOPIS, SFS and CURB-Side pickup orders
Q&A
Q: When testing a particular flow such as the 'Create order' flow, if we are not able to meet the target throughput with 1 instance and 1 thread, what should be our approach for increasing the number of threads or instances? Should we increase the number of threads first, or instances first before testing again?
A: Yes, you should first increase the threads and monitor for GC/CPU overhead. If GC/CPU is high, then scale the JVM instance. Also, please refer to our previous holiday readiness session where it covers the details on selecting the optimal performance profile in Next Generation environments.
A: Yes, you should first increase the threads and monitor for GC/CPU overhead. If GC/CPU is high, then scale the JVM instance. Also, please refer to our previous holiday readiness session where it covers the details on selecting the optimal performance profile in Next Generation environments.
Q: What is the optimal number of nodes to keep in a DG for sourcing?
A: Ideally not more than 20-30.
A: Ideally not more than 20-30.
Q: Why is performance tuning not implemented out-of-box on hosted environments? For example, configurations such as JMS session pooling and/or reuse connections are required for almost all clients who exhibit a higher order load.
A: JMS session pooling is enabled by default in OMoC. There are other JMS properties which are not enabled by default as some may not work effectively for all customers. However, we will be further improving in this area to add some additional general properties that we have observed work well for most customers.
A: JMS session pooling is enabled by default in OMoC. There are other JMS properties which are not enabled by default as some may not work effectively for all customers. However, we will be further improving in this area to add some additional general properties that we have observed work well for most customers.
Q: How should we select a profile (Balanced, Compute, etc.) for an agent/integration server? Is there a way to do performance testing around this to find the optimal profile?
A: We have documentation on how to test and determine the right profile and threads in the IBM Documentation.
A: We have documentation on how to test and determine the right profile and threads in the IBM Documentation.
Q: When all nodes have inventory in a group of 200 stores, how does smart sourcing help?
A: We filter by
1. The number of smart nodes
2. The latitude and longitude of the nodes with the ship address
3. The product availability date
A: We filter by
1. The number of smart nodes
2. The latitude and longitude of the nodes with the ship address
3. The product availability date
Q: When we have 30+ Stores for the SFS feature with region-based sourcing, will it have impact on sourcing? If so, is there a better way to source SFS orders when we have more than 30 stores?
A: Use smart sourcing
A: Use smart sourcing
Q: What kinds of database optimizations are required in order for inventory queries to run inventory calculations quickly? Is this something consultants need to triage, or is it handled by the IBM Order Management on Cloud team ?
A: Out-of-the-box queries are pre-tuned, but if you do observe a slow query, either when reviewing the SQLDEBUG log or the Self Service Database metric dashboard, feel free to raise a Support case. Additionally, you can use the DB Query Client to check for existing indices, as well as the developer toolkit database to run explain plans / db2 advise.
A: Out-of-the-box queries are pre-tuned, but if you do observe a slow query, either when reviewing the SQLDEBUG log or the Self Service Database metric dashboard, feel free to raise a Support case. Additionally, you can use the DB Query Client to check for existing indices, as well as the developer toolkit database to run explain plans / db2 advise.
Q: How many concurrent users are recommended for load testing scenarios?
A: At least 2x the number of expected users.
A: At least 2x the number of expected users.
Q: Following every performance test, extra data gets added to the database. For instance, consider that 30K orders are added to the database during a test. For the next performance test, these 30K will still be in the system. Should we have some cleanup strategy for this data generated during these tests? In the past, we would reset the database back to a prior state, from before the test had been executed, but it seems this is not possible with Order Management on Cloud. Are there any suggestions for this?
A: We do have an option to take a database backup using the Self Service Tool. This backup can later be used to restore the database to that prior state. Usually we recommend customers leverage purge agents to delete the data instead, but in the event incorrect data cannot be purged from the tables, you can use this Self Service Tool backup and restore strategy in the testing environment.
A: We do have an option to take a database backup using the Self Service Tool. This backup can later be used to restore the database to that prior state. Usually we recommend customers leverage purge agents to delete the data instead, but in the event incorrect data cannot be purged from the tables, you can use this Self Service Tool backup and restore strategy in the testing environment.
Q: During peak load testing, should Preprod's transactional data size be the same as in the Production environment?
A: It is better to have almost the same amount of transaction data in Preprod if you are trying to accurately see what might happen in Production during peak traffic.
A: It is better to have almost the same amount of transaction data in Preprod if you are trying to accurately see what might happen in Production during peak traffic.
Reading Material
The authors shared the following links to continue learning about the topic:
Presentation
[{"Type":"MASTER","Line of Business":{"code":"LOB59","label":"Sustainability Software"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SS6PEW","label":"Sterling Order Management"},"ARM Category":[{"code":"a8m0z000000cy01AAA","label":"Performance"}],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"All Versions"}]
Product Synonym
IBM Sterling Order Management, OMoC, Order Management on Cloud
Was this topic helpful?
Document Information
Modified date:
17 September 2021
UID
ibm16490339